The path to perfection: Quantum dots in electrically-controlled cavities yield bright, nearly identical photons

Optical quantum technologies are based on the interactions of atoms and photons at the single-particle level, and so require sources of single photons that are highly indistinguishable – that is, as identical as possible. Current single-photon sources using semiconductor quantum dots inserted into photonic structures produce photons that are ultrabright but have limited indistinguishability due to charge noise, which results in a fluctuating electric field. Conversely, parametric down conversion sources yield photons that while being highly indistinguishable have very low brightness. Recently, however, scientists at CNRS - Université Paris-Saclay, Marcoussis, France; Université Paris Diderot, Paris, France; University of Queensland, Brisbane, Australia; and Université Grenoble Alpes, CNRS, Institut Néel, Grenoble, France; have developed devices made of quantum dots in electrically-controlled cavities that provide large numbers of highly indistinguishable photons with strongly reduced charge noise that are 20 times brighter than any source of equal quality. The researchers state that by demonstrating efficient generation of a pure single photon with near-unity indistinguishability, their novel approach promises significant advances in optical quantum technology complexity and scalability.

Dr. Pascale Senellart and Phys.org discussed the paper, Near-optimal single-photon sources in the solid state, that she and her colleagues published in Nature Photonics, which reports the design and fabrication of the first optoelectronic devices made of quantum dots in electrically controlled cavities that provide bright source generating near-unity indistinguishability and pure single photons. "The ideal single photon source is a device that produces light pulses, each of them containing exactly one, and no more than one, photon. Moreover, all the photons should be identical in spatial shape, wavelength, polarization, and a spectrum that is the Fourier transform of its temporal profile," Senellart tells Phys.org. "As a result, to obtain near optimal single photon sources in an optoelectronic device, we had to solve many scientific and technological challenges, leading to an achievement that is the result of more than seven years of research."

While quantum dots can be considered artificial atoms that therefore emit photons one by one, she explains, due to the high refractive index of any semiconductor device, most single photons emitted by the quantum dot do not exit the semiconductor and therefore cannot be used. "We solved this problem by coupling the quantum dot to a microcavity in order to engineer the electromagnetic field around the emitter and force it to emit in a well-defined mode of the optical field," Senellart points out. "To do so, we need to position the quantum dot with nanometer-scale accuracy in the microcavity."

Senellart notes that this is the first challenge that the researchers had to address since targeting the issue of quantum dots growing with random spatial positions. "Our team solved this issue in 2008¹ by proposing a new technology, in-situ lithography, which allows measuring the quantum dot position optically and drawing a pillar cavity around it. With this technique, we can position a single quantum dot with 50 nm accuracy at the center of a micron-sized pillar." In these cavities, two distributed Bragg reflectors confine the optical field in the vertical direction, and the contrast of the index of refraction between the air and the semiconductor provides the lateral confinement of the light. "Prior to this technology, the fabrication yield of quantum dot cavity devices was in the 10^-4 – but today it is larger than 50%." The scientists used this technique to demonstrate the fabrication of bright single photon sources in 2013², showing that the device can generate light pulses containing a single photon with a probability of 80% – but while all photons had the same spatial shape and wavelength, they were not perfectly identical.

"Indeed, for the photons to be fully indistinguishable, the emitter should be highly isolated from any source of decoherence induced by the solid-state environment. However, our study showed that collisions of the carriers with phonons and fluctuation of charges around the quantum dot were the main limitations." To solve this problem, the scientists added an electrical control to the device³, such that the application of an electric field stabilized the charges around the quantum dot by sweeping out any free charge. This in turn removed the noise. Moreover, she adds, this electrical control allows tuning the quantum dot wavelength – a process that was previously done by increasing temperature at the expense of increasing vibration.

"I'd like to underline here that the technology described above is unique worldwide," Senellart stresses. "Our group is the only one with such full control of all of the quantum dot properties. That is, we control emission wavelength, emission lifetime and coupling to the environment, all in a fully deterministic and scalable way."

Specifically, implementing control of the charge environment for quantum dots in connected pillar cavities, and applying an electric field on a cavity structure optimally coupled to a quantum dot, required significant attention. "We had strong indications back in 2013 that the indistinguishability of our photons was limited by some charge fluctuations around the quantum dot: Even in the highest-quality semiconductors, charges bound to defects fluctuate and create a fluctuating electric field. In the meantime, several colleagues were observing very low charge noise in structures where an electric field was applied to the quantum dot – but this was not combined with a cavity structure." The challenge, Senellart explains, was to define a metallic contact on a microcavity (which is typically a cylinder with a diameter of 2-3 microns) without covering the pillar's top surface.

"We solved this problem by proposing a new kind of cavity – that is, we showed that we can actually connect the cylinder to a bigger frame using some one-dimensional bridges without modifying too much the confinement of the optical field." This geometry, which the researchers call connected pillars, allows having the same optical confinement as an isolated pillar while defining the metallic contact far from the pillar itself. Senellart says that the connected pillars geometry was the key to both controlling the quantum wavelength of dot and efficiently collecting its emission³.

In demonstrating the efficient generation of a pure single photon with near-unity indistinguishability, Senellart continues, the researchers had one last step – combining high photon extraction efficiency and perfect indistinguishability – which they did by implementing a resonant excitation scheme of the quantum dot. "In 2013, Prof. Chao-Yang Lu's team in Hefei, China showed that one could obtain photons with 96% indistinguishability by exciting the quantum dot state in a strictly resonant way⁴. Their result was beautiful, but again, not combined with an efficient extraction of the photons. The experimental challenge here is to suppress the scattered light from the laser and collect only the single photons radiated by the quantum dot."

Senellart adds that while removing scattered photons when transmitting light in processed microstructures is typically complicated, in their case this step was straightforward. "Because the quantum dot is inserted in a cavity, the probability of the incident laser light to interact with the quantum dot is actually very high. It turns out that we send only a few photons – that is, less than 10 – on the device to have the quantum dot emitting one photon. This beautiful efficiency, also demonstrated in the excitation process, which we report in another paper⁵, made this step quite easy."

The devices reported in the paper have a number of implications for future technologies, one being the ability to achieve strongly-reduced charge noise by applying an electrical bias. "Charge noise has been extensively investigated in quantum dot structures," Senellart says, "especially by Richard Warburton's group." Warburton and his team demonstrated that in the best quantum dot samples, the charge noise could take place on a time scale of few microseconds⁶ – which is actually very good, since the quantum dot emission lifetime is around 1 nanosecond. However, this was no longer the case in etched structures, where a strong charge noise is always measured on very short time scale – less than 1 ns – that prevents the photon from being indistinguishable. "I think the idea we had – that this problem would be solved by applying an electric field – was an important one," Senellart notes. "The time scale of this charge noise does not only determine the degree of indistinguishability of the photons, it also determines how many indistinguishable photon one can generate with the same device. Therefore, this number will determine the complexity of any quantum computation or simulation scheme one can implement." Senellart adds that in a follow-up study⁷ the scientists generated long streams of photons that can contain more than 200 being indistinguishable by more than 88%.

In addressing how these de novo devices may lead to new levels of complexity and scalability in optical quantum technologies, Senellart first discusses the historical sources used develop optical quantum technologies. She makes the point that all previous implementations of optical quantum simulation or computing have been implemented using Spontaneous Parametric Down Conversion (SPDC) sources, in which pairs of photons are generated by the nonlinear interaction of a laser on a nonlinear crystal, wherein one photon of the pair is detected to announce the presence of the other photon. This so-called heralded source can present strongly indistinguishable photons, but only at the cost of extremely low brightness. "Indeed, the difficulty here is that the one pulse does not contain a single pair only, but some of the time several pairs," Senellart explains. "To reduce the probability of having several pairs generated that would degrade the fidelity of a quantum simulation, calculation or the security of a quantum communication, the sources are strongly attenuated, to the point where the probability of having one pair in a pulse is below 1%. Nevertheless, with these sources, the quantum optics community has demonstrated many beautiful proofs of concept of optical quantum technologies, including long-distance teleportation, quantum computing of simple chemical or physical systems, and quantum simulations like BosonSampling." (A BosonSampling device is a quantum machine expected to perform tasks intractable for a classical computer, yet requiring minimal non-classical resources compared to full-scale quantum computers.) "Yet, the low efficiency of these sources limits the manipulation to low photon numbers: It takes typically hundreds of hours to manipulate three photons, and the measurement time increases exponentially with the number of photons. Obviously, with the possibility to generate more many indistinguishable photons with an efficiency more than one order of magnitude greater than SPDC sources, our devices have the potential to bring optical quantum technologies to a whole new level."

Other potential applications of the newly-demonstrated devices will focus on meeting near-future challenges in optical quantum technologies, including scalability of photonic quantum computers and intermediate quantum computing tasks. "The sources presented here can be used immediately to implement quantum computing and intermediate quantum computing tasks. Actually, very recently – in the first demonstration of the superiority of our new single photon sources – our colleagues in Brisbane made use of such bright indistinguishable quantum dot-based single photon sources to demonstrate a three photon BosonSampling experiment⁸, where the solid-state multi-photon source was one to two orders-of-magnitude more efficient than downconversion sources, allowing to complete the experiment faster than those performed with SPDC sources. Moreover, this is a first step; we'll progressively increase the number of manipulated photons, in both quantum simulation and quantum computing tasks."

Another target area is quantum communications transfer rate. "Such bright single photon sources could also drastically change the rate of quantum communication protocols that are currently using attenuated laser sources or SPDC sources. Yet, right now, our sources operate at 930 nm when 1.3 µm or 1.55 µm sources are needed for long distance communications. Our technique can be transferred to the 1.3 µm range, a range at which single photon emission has been successfully demonstrated – in particular by the Toshiba research group – slightly changing the quantum dot material. Reaching the 1.55 µm range will be more challenging using quantum dots, as it appears that the single photon emission is difficult to obtain at this wavelength. Nevertheless, there's a very promising alternative possibility: the use of a 900 nm bright source, like the one we report here, to perform quantum frequency conversion of the single photons. Such efficient frequency conversion of single photons has recently been demonstrated, for example, in the lab of Prof. Yoshie Yamamoto at Stanford⁹."

Regarding future research, Senellart says "There are many things to do from this point. On the technology side, we will try to improve our devices by further increasing the source brightness. For that, a new excitation scheme will be implemented to excite the device from the side, as was done by Prof. Valia Voliotis and her colleagues on the Nanostructures and Quantum Systems team at Pierre and Marie Curie University in Paris and Prof. Glenn Solomon's group at the National Institute of Standards and Technology (NIST) in Gaithersburg, Maryland. Applying this technique to our cavities should allow gaining another factor of four on source brightness. In addition, operating at another wavelength would be another important feature for our devices, since as discussed above, this would allow using the source for quantum telecommunication. For example, a shorter wavelength, in the visible/near infrared range, would open new possibilities to interconnect various quantum systems, including ions or atoms through their interaction with photons, as well as applications in quantum imaging and related fields."

The researchers also want to profit from the full potential of these sources and head to high photon number manipulation in, for instance, quantum simulation schemes. "We're aiming at performing BosonSampling measurements with 20-30 photons, with the objective of testing the extended Church Turing thesis and proving the superiority of a quantum computer over a classical one." The original Church Turing thesis, based on investigations of Alonzo Church and Alan Turing into computable functions, states that, ignoring resource limitations, a function on the natural numbers is computable by a human being following an algorithm, if and only if it is computable by a Turing machine.

Another promising impact on future optical quantum technologies is the generation of entangled photon pairs. "A quantum dot can also generate entangled photon pairs, and in 2010 we demonstrated that we could use the in situ lithography to obtain the brightest source of entangled photon pairs¹⁰. That being said, photon indistinguishability needs to be combined with high pair brightness – and this is the next challenge we plan to tackle. Such a device would play an important role in developing quantum relays for long distance communication and quantum computing tasks."

Senellart tells Phys.org that other areas of research might well benefit from their findings, in that devices similar to the one the scientists developed to fabricate single photon sources could also provide nonlinearities at the low photon count scale. This capability could in turn allow the implementation of deterministic quantum gates, a new optical quantum computing paradigm in which reversible quantum logic gates – for example, Toffoli or CNOT (controlled NOT) gates– can simulate irreversible classical logic gates, thereby allowing quantum computers to perform any computation which can be performed by a classical deterministic computer. "Single photons can also be used to probe the mechanical modes of mechanical resonator and develop quantum sensing with macroscopic objects. Other applications," she concludes, "could benefit from the possibility to have very efficient single photon sources, such as an imaging system with single photon sources that could allow dramatically increased imaging sensitivity. Such technique could have applications in biology where the lower the photon flux, the better for exploring in vivo samples."

© 2016 Phys.org